Blog

I recap my experience at New Builds 2024 in Toronto, Canada, and share how my team built a generative choose-your-own-adventure game in 24 hours.

I explain the key ideas of the paper "Denoising Diffusion Probabilistic Models" by Ho et al., which introduced the concept of diffusion models for image generation.

I provide a small example on how to ingest, chunk, and embed 10-K filings into a Supabase database, and then use LangSmith to evaluate two models for a retrieval task.

I provide a brief overview of LangChain, an all-in-one LLM development stack, from prototyping to monitoring. A gentle introduction to RAG is also included.

I show how to fine-tune GPT-2 for a classification task using Hugging Face Transformers.

I provide a detailed guide on how to build a multi-layer LSTM from scratch in Python.

This post demonstrates how to use GPT-4 Vision API automatically generate advertising copy from just product images using the OpenAI API.

I explain the architecture of the Transformer model in detail, including the encoder and decoder blocks, and the self-attention mechanism

I build a model using regression techniques to predict IMDb ratings of new movies based on their features, and analyze the importance of each feature in determining the rating

I analyze the performance of teams and players in Group D of the ESL Pro League Season 14, highlighting key metrics, map distribution, and standout players from the group stage